Re: [PATCH] firewire: core: use non-reentrant workqueue where necessary
From: Tejun Heo
Date: Mon Oct 11 2010 - 09:29:55 EST
Hello, Stefan.
On 10/10/2010 04:55 PM, Stefan Richter wrote:
...
> Therefore queue all of these types of work items on system_nrt_wq
> instead of system_wq. The former guarantees non-reentrance across all
> CPUs, the latter only on the CPU which schedules a work item.
>
> As a bonus, other subsystems which flush system_wq won't be held up if
> the firewire subsystem spends a lot of time in an extraordinarily long
> fw_device.work.
Awesome, thanks a lot for doing this. BTW, I'm in the process of
hunting down all flush_scheduled_work() users and plan to remove
workqueue flushing capability from all system workqueues.
> Signed-off-by: Stefan Richter <stefanr@xxxxxxxxxxxxxxxxx>
For the workqueue part,
Acked-by: Tejun Heo <tj@xxxxxxxxxx>
--
tejun
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/